Python Basemap drawgreatcircle function

Question

Hi I'm trying to map a graph of ip addresses onto a world map using the greatcircle function in Basemap (part of matplot lib) but every time I connect two points that span across the Pacific Ocean (i.e. Somewhere in the USA west coast and australia) there is a horizontal line that shows up in my plot.

I know it's because of this problem:

Note Cannot handle situations in which the great circle intersects the edge of the map projection domain, and then re-enters the domain. (http://matplotlib.org/basemap/api/basemap_api.html#mpl_toolkits.basemap.Basemap.drawgreatcircle)

Just wondering if anyone knew how to fix it or knew of another package which didnt' have this problem. Thanks!

Answer 1

To solve this in Basemap for specific cases you can check the differences of the x vertices in the resulting path. From there, we can cut the bad path into two sections. I have thrown together my own example of doing this with basemap:

from mpl_toolkits.basemap import Basemap
import numpy as np
import matplotlib.pyplot as plt


m = Basemap(projection='cyl', lon_0=0, resolution='c')

m.fillcontinents(color='coral',lake_color='aqua')
m.drawmapboundary(fill_color='aqua')

places = {'Mexico City': (19.05, -99.366667),
          'London': (51.507222, -0.1275),
          'Sydney': (-33.859972, 151.211111),
          'Cape Town': (-33.925278, 18.423889),
          'Delhi': (28.61, 77.23),
          }

network = [
           ('London', 'Delhi'),
           ('Mexico City', 'Sydney'),
           ]


for source, target in network:
    lat1, lon1 = places[source]
    lat2, lon2 = places[target]
    line, = m.drawgreatcircle(lon1, lat1, lon2, lat2, lw=3)

    p = line.get_path()
    # find the index which crosses the dateline (the delta is large)
    cut_point = np.where(np.abs(np.diff(p.vertices[:, 0])) > 200)[0]
    if cut_point:
        cut_point = cut_point[0]

        # create new vertices with a nan inbetween and set those as the path's vertices
        new_verts = np.concatenate(
                                   [p.vertices[:cut_point, :], 
                                    [[np.nan, np.nan]], 
                                    p.vertices[cut_point+1:, :]]
                                   )
        p.codes = None
        p.vertices = new_verts


plt.show()

The result:

This is not a general solution, and will only work when you have latitudes and longitudes. Solving the general problem is much, much harder and is the focus of cartopy (http://scitools.org.uk/cartopy/docs/latest/). I will post an example of doing this the exact same plot in cartopy in the next couple of days.

HTH

Answer 2

Another package which doesn't have this problem: cartopy (as yet, unannounced). The example found here is relevant.

HTH